Hopkins scientists overcome main obstacle to making tons of short, drug-like proteins

Two Johns Hopkins scientists have figured out a simple way to make millions upon millions of drug-like peptides quickly and efficiently, overcoming a major hurdle to creating and screening huge “libraries” of these super-short proteins for use in drug development.

“Our work dramatically increases the complexity of peptide libraries that can be created and the speed with which they can be made and processed,” says Chuck Merryman, Ph.D., a postdoctoral fellow who developed the new technique. “In an afternoon, we’ll be able to make literally millions of millions of different peptides with medicinal potential.”

Usually less than 40 building blocks long, peptides act as important messengers and hormones in the body. But because their building blocks, called amino acids, are quickly recycled, peptides made from the 20 naturally occurring amino acids don’t last long enough to be useful as medicines. However, adding a tiny methyl group to each amino acid gives the resulting peptide “drug-like” stability.

Writing in the April 19 issue of Chemistry & Biology, the Hopkins scientists reveal that using a simple chemical reaction, first reported in the early 1980s, allows them to convert en masse the naturally occurring amino acids to ones that form more stable peptides.

The tricky part, Merryman says, was figuring out how to do the conversion while the amino acids were attached to transfer RNA, a carrier molecule required for the biological production of peptides. The advance makes it possible to build upwards of 10,000,000,000,000 — that’s 1 with 13 zeros behind it — stabilized, 10-block-long peptides at once.

“The idea of creating large peptide libraries and testing them for medicinal uses has been around a long time, but until now it’s just not been very practical,” says Merryman.

A key aspect of all scientists’ efforts to create libraries of drug-like peptides is “biology in a dish” — harnessing the same machinery cells use to read genetic instructions and assemble correct proteins. Since at least the 1970s, scientists have known that this machinery, called the ribosome, also can string together a wide variety of artificial amino acids, as long as the fake building block is tied to transfer RNA that the ribosome can use to “decode” genetic information.

“There are a number of steps to the process of building peptides, natural or not, and each one has created problems for building large libraries of random drug-like peptides,” says Merryman.

A complex of RNA and proteins, the ribosome “reads” three-bit sections of messenger RNA and recruits a complementary three-bit-containing piece of transfer RNA, which is attached to its corresponding amino acid. The ribosome’s machinery then chops off the amino acid and adds it to the growing peptide string.

To harness this natural process to do their bidding, scientists have tried to make various artificial amino acids attached to transfer RNA, and to have the ribosome use those novel components while reading genetic instructions, the messenger RNA.

“There’s been some success, but no one’s been able to do this with multiple artificial amino acids at once or to create very large numbers of peptides that are entirely artificial,” says Merryman.

Merryman starts with a mixture of the 20 naturally occurring amino acids, already tethered to their transfer RNA sequences. In the new process, a first chemical step temporarily protects one reactive side of the exposed nitrogen atom of the amino acid, and a second step adds the methyl group to the nitrogen’s other open spot. The final step uses ultraviolet light to remove the protecting group added in step one.

The result is a single pot of the 20 natural amino acids, still attached to the appropriate tRNAs, but 19 of them now with that all-important methyl group. (One amino acid, proline, remains unchanged — once its nitrogen is protected, there’s no room for the methyl group.)

“It’s pretty simple chemistry, and it’s kind of amazing it hadn’t already been applied to this problem,” says Merryman. “The process gave us efficient and essentially complete conversion to the modified amino acids. Since the process works the same for all amino acids, you don’t have to treat each one separately and then mix them at the end, which speeds things up considerably.”

Merryman and mentor Rachel Green, Ph.D., an associate professor of molecular biology and genetics and an associate investigator of the Howard Hughes Medical Institute, have a patent on the synthetic process under review at the U.S. Patent and Trademark Office.

To make a peptide library from the pot of modified amino acids, researchers would mix the pot with ribosomes and random sequence messenger RNAs that reflect the possible combinations of the 20 artificial amino acids for the desired length — say, 10 blocks long. The ribosomes then churn away, making the peptides.

To search for potential medicines, the peptide library would be mixed with a molecule of interest, say Hedgehog, a protein implicated in cancer. Peptides that bind Hedgehog would stick, and those that don’t would be washed away. Then, using a messenger RNA identifier Merryman has engineered to stay attached to each peptide, the genetic instructions for “winning” peptides can be selectively amplified. By repeatedly building the library, testing the peptides, and amplifying the winning ones, the library “evolves,” gradually accumulating the most promising peptides.

The research was funded by the National Institutes of Health and the Howard Hughes Medical Institute.